An inexpensive antigen for serodiagnosisof Chagas’ disease.

Luis Briceño1, Eva Mary Rodríguez2, Mehudy Medina3, Yelitza Campos1, Walter Mosca1,

Amanda Briceño4 and Graciela León5.

1Laboratory of Physiopathology, Institute of Biomedicine, Faculty of Medicine,Central University of Venezuela, 2Research Center in Public Health “Jacinto Convit”,Faculty of Medicine, Central University of Venezuela, Direccion General de SaludAmbiental, 3Ministerio de Salud, 4Universita degli Studi di Milano, Italia and5Municipal Blood Bank, Caracas, Venezuela.

Key words: Chagas’disease, diagnostic, antigen, inexpensive.

Abstract. In this prospective study we evaluated the performance charac-teristics of a specific and sensitive antigen preparation (AgA) used in an en-zyme-linked immunosorbent assay (ELISA) for the detection of anti-Trypanosoma cruzi antibodies in serum samples, for Chagas’ disease diagno-sis. The antigen production was achieved by combination of nutritional stressand autoclaving the parasites. Specificity and sensitivity were evaluated in twoseparate tests, using 152 sera from healthy individuals and 175 sera from Cha-gas’ patients (70 by xenodiagnosis). Cross-reactivity was tested using 289 serafrom patients who had a parasitological diagnosis of a disease known to in-duce antigenic responses towards T. cruzi. All of these sera were tested withour AgA-ELISA and with 3 commercial diagnosis kits. To evaluate the agree-ment of results between our AgA-ELISA and a “gold standard” test for Chagas,we tested 566 sera from an endemic area. Results: sensitivity and specificitywere 100%; cross-reactivity was the lowest compared with commercial kits.Overall agreement with the gold standard test was excellent (kappa=0.92).AgA-ELISA exhibits levels of sensitivity, specificity and cross-reactivity compa-rable or superior to those shown, obtained with the commercial kits used inour country, while being at least 10 times less expensive. This balance be-tween diagnostic accuracy and cost makes AgA-ELISA useful for blood bankscreening in poor regions of the world suffering from Chagas’ disease. Furthervalidations of this antigenic formulation in other countries are necessary.

Invest Clin 51(1): 101 - 113, 2010

Corresponding author: Luis Briceño. Instituto de Biomedicina, San Nicolás a Providencia, San José. Caracas,Venezuela. Phone/fax: (058 212) 862 20 77. E-mail: lbricenozoppi@gmail.com.

Un antígeno económico para el serodiagnóstico de la enfermedadde Chagas.Invest Clin 2010; 51(1): 101 - 113

Palabras clave: Enfermedad de Chagas, serodiagnóstico, antígeno, económico.

Resumen. Este estudio fue realizado para evaluar las características desensibilidad y especificidad de una formulación antigénica (AgA), producida abajo costo, para detectar anticuerpos dirigidos a Trypanosoma cruzi, enmuestras de suero de pacientes con enfermedad de Chagas. El AgA fue produ-cido por el efecto combinado de estrés nutricional y autoclave de los parási-tos. La especificidad y sensibilidad fueron evaluadas en dos estudios separa-dos, con 152 sueros de individuos sanos y 175 de pacientes Chagásicos. Lareactividad cruzada con 289 sueros de pacientes con diagnóstico parasitológi-co de enfermedades con anticuerpos que reaccionan con antígenos de T. cru-

zi. Estos sueros fueron evaluados con AgA-ELISA y con tres estuches comer-ciales. 566 muestras de suero provenientes de un área endémica, fueron em-pleadas para estudiar la concordancia entre nuestro diagnostico y una pruebadesignada por nosotros como patrón oro estándar. Resultados: la sensibilidady especificidad fue de 100%. El AgA presento el más bajo porcentaje de reacti-vidad cruzada, respecto a los estuches comerciales evaluados. La concordan-cia con la prueba patrón oro, en Venezuela, fue excelente (kappa=0,92). Enconclusión, Aga-ELISA, presentó niveles de sensibilidad, especificidad y dereactividad cruzada, comparables o superiores a los obtenidos por los tres es-tuches comerciales mas empleados en el país, pero con costos de producciónal menos 10 veces menor. Este balance conveniente, favorece su potencial usopara el despiste en los bancos de sangre de los países pobres y endémicos parala enfermedad de Chagas. Futuras validaciones de esta formulación en otrospaíses es necesaria.

Received: 01-07-09. Accepted: 10-12-2009.

INTRODUCTION

Chagas’ disease is an important publichealth problem; according to the WorldHealth Organization (WHO) nearly 15 mil-lion people are infected and some 28 mil-lion are at risk of acquiring the disease (1).In Latin America, considering the numberof disability-adjusted life years lost due toChagas’disease, it is the third largest tropi-cal disease burden after malaria and schis-tosomiasis (1). Even though the Southern

Cone Initiative to Control/Eliminate Cha-gas Disease (INCOSUR) has made impor-tant progress (2), a large proportion ofLatin American population is exposed tothe parasite. For this reason, strategy forChagas’ disease eradication over the period1996-2010 is based on interrupting vecto-rial transmission and on the systematicscreening of blood donors (3).

Parasitological diagnosis during thechronic phase of the illness has levels ofsensitivity that oscillate among 50-65%,

Investigación Clínica 51(1): 2010

102 Briceño y col.

with time and effort consuming techniques(4, 5). Consequently, serological diagnosisis the most common way to determine thepresence of anti-T. cruzi antibodies. TheWHO recommends (6), and in some coun-tries, like Brazil, it is mandatory (7), tocarry out this screening with at least twodifferent tests, because antigens usuallypresent important problems due to false ordiscrepant results (8). In order to meet thisobjective, efforts have been made to substi-tute the whole parasite by purified parasiteproteins (9), but their purification is ardu-ous and the yield is very low due to proteas-es (10). Cloning and expressing the recom-binant antigens have given excellent resultswith increasing specificity and sensitivity(11-15). However, although most of thesepurified proteins have been successfullytested in research laboratories, they are notcommercially available or have not been in-cluded in blood banks screenings for costor/and technical reasons (10.11). Due tothese difficulties, together with financialreasons, in countries where this disease isendemic, the diagnosis is made using a sin-gle serological test.

Considering the above factors, the de-velopment of an economic kit in an En-zyme-Linked Innunosorbent Assay (ELISA)format has become necessary. As the com-ponents of the ELISA are relatively com-mon, the goal should be to produce a lowcost antigen. We have recently shown (16)that combined treatment (nutrient stressand autoclaving) of epimastigote forms en-hance the sensitivity and reduce cross-reac-tivity. The results of this study allow us toobtain an antigen that could be used in theserological diagnosis.

The aim of the present investigationwas to evaluate and validate a serologicalELISA developed with an antigenic formula-tion with high sensitivity, specificity andlow cost, so that it can be used as seconddiagnostic test for Chagas’ disease.

MATERIALS AND METHODS

Antigen

The antigen A (AgA) was prepared byautoclaving and nutrient stress ofepimastigote forms in stationary phaseTrypanosoma cruzi, Y strain, as previouslydescribed (16). Briefly, parasites weregrown in liver infusion tryptose (LIT) me-dium (17), diluted 50:50 with RPMI me-dium (Gibco) and supplemented with 5% fe-tal bovine serum. Once the epimastigoteforms culture reached the stationary phaseof growth, the parasites were processed toeliminate the culture media by centri-fugation at 2000 g for 10 min, in PBS, pH7.2. The parasites were resuspended inDMEM supplemented with gentamicin (50µg/mL) and incubated at 25°C. After 24hours in the above conditions, the culturewas centrifuged at 2000g for 10 min, thesupernatant discarded and the parasiteswere resuspended in sterile PBS pH 7.2.The last step was repeated once more. Afterthe last wash, they were resuspended at 30× 106 parasites/mL, and autoclaved (15 lb,15 min). After autoclaving, the suspensionlook cloudy (Proteins and remnants of theparasite) is named by us as AgA, and keptat 4°C until used.

A simple reversed-phase high-perfor-mance liquid chromatographic (VYDACC18 column) method was developed andvalidated to evaluate the batches produc-tion variation. The batches evaluatedfor the antigen formulation was highly re-producible.

Western blot

The AgA suspension (one or threetimes concentrated) and epimastigoteforms in stationary phase (30 × 106) wasseparated by SDS-PAGE, under denaturingand reducing condition, at the same para-site concentration, transferred to nitro-

Vol. 51(1): 101 - 113, 2010

An inexpensive antigen for serodiagnosis of Chagas’ disease 103

cellulose sheet as previously described (18).The sheet was blocked with 150 mM NaCl,50 mM tris-base (TBS), 5% low fat milk, for1 h at room temperature. To study the ef-fect of increasing temperature on the crossreactivity, the sheet was incubated with apool of six mucocutaneous Leishmanisis pa-tients’ sera and membrane with AgA (3X)strips were incubated with a pool of Chagaspatients’ sera, with low, medium or high ti-ter to T cruzi. For all test, the membraneswere incubated with sera (1:2000) dilutedin TBS, for 1 h, at room temperature,washed, and the bound antibodies were de-tected with biotinylated goat anti-humanIgG (GIBCO Laboratories, Grand Island,NY, USA) for 30 min washed three timeswith TBS-T, and incubated with avidin-alka-line phosphatase. The presence of boundantibodies was detected with BCIP/NBT(GIBCO Laboratories, Grand Island, NY,USA) and analyzed using a Gel Doc 2000System (BIO-RAD, USA).

Sera samples

All the sera samples used in this studywere taken under conditions, proceduresand use approved by the ethics committeeof the Biomedicine Institute, UCV.

Pilot reactivity test

To demonstrate the capability of ourAgA-ELISA to discriminate between healthycontrols and Chagas’ disease patients, weused a total of 92 sera samples from healthyindividuals and 106 sera samples fromchagasic patients (positive to lymphoblasto-genesis and in at least two serological as-says), 37 of which were from chronicchagasic cardiomyopathy (CCC) and 68from asymptomatic (ASY) patients. Thiswould give us additional information aboutthe ability of the test to differentiate be-tween CCC and ASY patients.

Sensitivity and specificity test

To evaluate the sensitivity and specific-ity of the AgA-ELISA we used 60 sera fromhealthy individuals (with negative results inthree different serological tests) and 70previously positive sera from chagasic pa-tients (positive by xenodiagnosis). Sampleswere assayed using a single blind model.

Cross reactivity test

In the areas where Chagas disease isendemic, there is often a superposition ofparasitic diseases, the antibodies of whichhave shown to have cross reactivity with T

cruzi. To determine the cross reactivity ofthe AgA-ELISA, we used 289 sera fromparasitologically diagnosed patients withother diseases known to induce antigenicresponse against T. cruzi: 89 with cutane-ous Leishmaniasis (56 localized cutaneous,LCL; 12 mucocutaneous, MCL and 21diffusex cutaneous, DCL), and 111 with vis-ceral Leishmaniasis (VL), and sera from pa-tients infected with Toxoplasmosis (15),Schistosomiasis (39), Malaria (30) andGiardiasis (5). All sera were also independ-ently tested by the Caracas Municipal BloodBank, using three of the most commonlycommercial kits used in the country(Bioschile Ingenieria Genetica S.A Kit,BioKIT and Pharmatest (LaboratoriosPharmatest).

As we do not have an experimentaltool to confirm mixed infections, to com-pare the serological cross-reactivity be-tween different tests studied, we excludedall samples that came up positive in all thefour Chagas tests used (3 commercial kitsplus our AgA-ELISA).

Tests in sera from endemic areas.

Accuracy and reliability

To evaluate the agreement of resultsbetween our AgA-ELISA and a “gold stan-dard” test for Chagas’ disease diagnosis, we

Investigación Clínica 51(1): 2010

104 Briceño y col.

tested 566 serum samples from an endemicarea. These samples were tested simulta-neously and independently by our labora-tory and by the laboratory of the epidemio-logical division of the Health and Social De-velopment Ministry (MSDS, in Spanish). InVenezuela we do not have a real gold stan-dard test but the epidemiology division of theMSDS is the oldest and the most experiencedstate laboratory for Chagas’ disease diagno-sis. The MSDS diagnostic was obtained by theconsensus results of three different methods:epimastigote immunofluorescence, indirecthemagglutination and ELISA.

AgA-ELISA

Chagas’antibody detection was donethrough a previously described indirectmethod (19). A dilution (in buffer carbon-ate-bicarbonate) the AgA suspension wasused to sensitized 96-well plates at a con-centration 0.8 µg /well (protein on suspen-sion). Briefly, the plates (Maxi-Sorp, Nunc,Denmark) were sensitized with the AgA, at4°C, in sodium carbonate-bicarbonatebuffer (0.05 M, pH 9.6). After overnight in-cubation the plate was blocked for onehour, with 1% fetal bovine serum and 4%skim milk (Nestlé) in Tris-buffered salineplus Tween 20 (10mM Tris pH 7.5, 100mMNaCl, 0.1% Tween 20).The plates were thenincubated with human sera (1:100), for 30min, at 37°C. As second antibody we usedalkaline phosphatase conjugated goat anti-human IgG (GIBCO) at a 1: 2000 dilution,for 30 min. The immune complexes weredeveloped by the addition of 100 µL/ well ofp-nitrophenyl phosphate as substrate, indiethanolamine buffer assay reagent kit(Pierce, Co) and the 405 nm absorbancewas determined using Multiskan Ex,Labsystems.

Assay cut off

Cut off of the AgA-ELISA was appliedas follows:

Cut off = Y + Z

In each assay we included a samplefrom a serum pool from 46 healthy personsin four of the ELISA plate wells and we cal-culated the mean, which we called Z.

Y represents the individual dispersionvalue from the mean value obtained fromthe same 46 serums. We calculated thestandard deviation (DESVest) for each se-rum of this pool. Then we calculated thepercentage of deviation from the mean(P.DESVest). This value is used for all thedeterminations from this pool. Then,

Y = [(P.DESVest × Z)/ 100] × 2

Data analysis

Data were analyzed with the GraphPadPrism 3.0 (Graphpad Software, Inc, SanDiego, CA). To assess the validity underfield conditions of the diagnostic test, wecalculated the descriptive statistics (sensi-tivity, specificity, efficiency rate and likeli-hood ratios) with their 95% CIs. To evaluatethe concordance between our test and theepidemiological division diagnosis, we cal-culated Cohen’s kappa statistics (20). Toassess agreement beyond that expected bychance with their respective 95% confi-dence intervals, we also calculated preva-lence and bias adjusted kappa (PABAK)(21). The agreement degree was defined asfollows: low agreement � 0; slight from0.01 to 0.20; fair from 0.21-0.40; moderatefrom 0.41-0.60; substantial from 0.61 to0.80 and almost perfect of 0.81-1.00 (22).For the analysis we used the STATA Release9.2® software (Stata, College Station, Tex),and DAGstat spreadsheet (23).

RESULTS

AgA characterization

Because T. cruzi shares antigenic de-terminants with other pathogens and par-

Vol. 51(1): 101 - 113, 2010

An inexpensive antigen for serodiagnosis of Chagas’ disease 105

ticularly with members of the Leishmania

genus, many serodiagnosis tests for theChagas disease have a persistent highnumber of false positive results. This crossreactivity may be lost through heat treat-ment, thus changing their conformationshape or loosing the thermolabile anti-gens. In this sense, we have studied this ef-fect in epimastigote forms of Trypanosoma

cruzi, in stationary phase, by increasing thetemperature or autoclaving. In Fig. 1, wefound that the reactivity of pooled serafrom patients with Leishmaniasis diseasedecreases with increasing temperature andthe best result was obtained with auto-claving. All bands were not completely “de-stroyed” by the last treatment, because thethree times concentrated sample was recog-nized by antibodies. Recently (16), we haveobserved (immunofluorescence) that pa-rameters, like culture conditions (parasitesin stationary phase of growth plus nutri-tional stress), significantly affect sensitivity,and induce changes in the distribution anddensity of antigens recognized by a pool ofsera from experimentally infected mice. Inthe same study, we also tested whetherwhole parasites extract or sub fractions im-prove sensitivity or cross-reactivity whenthey are autoclaved. Surprisingly, the bestsensitivity and lowest cross-reactivity wereobserved with whole parasite extract fromcultures under nutrient stress and auto-claved.

As we identified the conditions for ob-taining a low cost antigen, we studied by aWestern Blot assay how the AgA formula-tion was recognized by a pool of Chagas pa-tients’ sera, with low, medium or high titersto T. cruzi (Fig. 2). The AgA was recognizedfor all conditions evaluated. Once demon-strated by a Western Blot assay that theAgA formulation was recognized by antibod-ies from the chagasic patients, its feasibilityas diagnostic test was studied with a groupof sera from healthy subjects, ASY and CCC

patients. Results are shown in Fig. 3. Wefound that the AgA-ELISA was capable ofrecognizing 100% of sera from infected pa-tients. We did not find any differences inantibody levels between ASY patients andthe ones with CCC.

Sensitivity and specificity Test

Using a blind assay to test 130 sera(70 positive by xenodiagnostic), we con-firmed 100% of specificity and sensitivity ofthe AgA-ELISA.

Estimation of cross reactivity

When AgA-ELISA was tested using serafrom patients diagnosed with other para-

Investigación Clínica 51(1): 2010

106 Briceño y col.

1 2 3 4 5 6

200

110

97

72

60

45

32

MW

Fig. 1. Increases in temperature and its effecton cross-reactivity with T. cruzi. Evalu-ating the heat treatment on across arange of temperatures: 70°C (lane 1),80°C (lane 2), 90°C (lane 3) and 100°C(lane 4), for 15 min, at the same para-site concentration (30 × 106/mL). TheAgA, was studied at the same concentra-tion, 30 × 106/mL (lane 6) or treetimes concentrated (lane 5). Thenitrocellulose sheet was incubated witha pool the six sera (1/2000) from pa-tients with Leishmaniasis Americana.MW: molecular weight marker.

sitic diseases, we found a low percentage ofcross reactivity. There was 8.93% (5/56) ofreactivity with sera from patients with LCL;16.66% (2/12) with MCL and 4.76% (1/21)with sera of patients with DCL (Fig. 4). The8 sera that showed cross reactivity for T.

cruzi were re-tested using commercial kitswhich are used frequently in blood banksfor diagnosis in Venezuela. Six sera werepositive with all commercial kits, suggest-ing a mixed infection.

Fig. 5 shows results obtained with 111serums from VL patients; twenty-four ofthem were from a presumably non-endemicarea for Chagas disease (Nueva Esparta

Vol. 51(1): 101 - 113, 2010

An inexpensive antigen for serodiagnosis of Chagas’ disease 107

L M H200

110

97

72

60

45

32

MW

Fig. 2. AgA, western blot analysis of antibodiesto T. cruzi. SDS-PAGE of AgA (3X) pro-teins were separated by electrophoresison a 10% gel and electrotransferredonto nitrocellulose membrane. AgAstrips were incubated with a pool (sixsera for each condition) of Chagas pa-tients’ sera with low (L), medium (M)or high (H) titer to T. cruzi. For all testthe membranes were incubated withsera (1:2000) diluted in TBS. MW: mo-lecular weight mark.

Healthy ASY CCC

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

OD

40

5

Fig. 3. Pilot Reactivity Test: Distribution of in-dividual OD405 values obtained byAgA-ELISA (IgG antibodies against T.

cruzi proteins), in 92 sera samples fromhealthy individuals and 106 sera sam-ples from chagasic patients. Dotted linedenotes cut-off. ASY: Chagasic asymp-tomatic patients (69 sera); CCC: chronicchagasic cardiomyopathy (37 sera).

LCL MCL DCL

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

OD

405

Fig. 4. Cross reaction with sera from patientswith cutaneous Leishmaniasis. Distribu-tion of individual OD405 values obtainedby AgA-ELISA (IgG antibodies against T.

cruzi proteins), in 89 sera from patientswith cutaneous Leishmaniasis. Local-ized cutaneous Leishmaniasis (LCL, 56sera). Mucocutaneous Leishmaniasis(MCL, 12 sera). Diffused cutaneousLeishmaniasis (DCL, 21 sera). Dottedline denotes cut-off.

State). We did not find cross reactivity insamples from the non-endemic area, but inthe other 87 serum samples we observed a9.20% (8/87) of positives. All 8 samples re-sulted positive with at least one of the com-mercial diagnostic kits for Chagas disease:Bioschile Ingenieria Genetica S.A., BioKITand Pharmatest, Laboratorios Pharmatest.

Additionally, 89 sera from patients in-fected with Toxoplasmosis (15), Schistoso-miasis (39), Malaria (30) and Giardiasis (5)were evaluated. Of these only one serumfrom a patient with Schistosomiasis gave apositive result (Fig. 6).

We compared our AgA-ELISA withthree commercial kits for Chagas’ diseasediagnosis (Table I). With our formulationwe obtained the lowest cross reactivity re-sults. The higher percentage of cross reac-tivity was observed in VL patients.

Agreement evaluation

Five hundred and sixty-six sera (89positive sera and 477 negative sera diag-nosed by the MSDS) were simultaneouslyand independently tested in our lab, usingAgA-ELISA, and by the epidemiological divi-sion of MSDS, using 3 different methods(epimastigotes immunofluorescence, indi-rect hemagglutination and ELISA) for a sin-gle result. In this field evaluation, theMSDS results were considered as referencetests. Results are shown in Table II. Afteranalyzing the results, we found good agree-ment for all parameters evaluated: 98.41%(95% CI: 97.60-99.21) for negative agree-ment and 92.15% (95% CI: 88.18-96.11) forpositive agreement. Cohen’s kappa coeffi-cient was 90.56% (95% CI: 85.87-95.26)and PABAK was 94.70%. The correct classi-fication rate (Efficiency) was 97.35% (95%CI: 95.67- 98.51). The sensitivity and speci-ficity index were 98.88% and 97.06% re-spectively. The likelihood ratio of negativetests (86.39) was higher than the likelihoodratio of positive tests (33.69). Likelihood

Investigación Clínica 51(1): 2010

108 Briceño y col.

Endemic No endemic0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

OD

405

Fig. 5. Cross reaction with sera from patientswith visceral Leishmaniasis (VL). Distri-bution of individual OD405 values ob-tained by AgA-ELISA (IgG antibodiesagainst T. cruzi proteins), with 111 pa-tients sera. Twenty-four of them werefrom a presumably non-endemic areafor Chagas disease (Nueva EspartaState) and 87 sera from chagas endemicarea. Dotted line denotes cut-off.

Toxo Shisto Giardia Malaria

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

DO

405

nm

Fig. 6. Cross reaction with sera from patientswith Toxoplasmosis (Toxo, 15 sera),Schistosomiasis (Shisto, 39 sera),Giardiasis (Giardia, 5 sera) and Malaria(30 sera). Distribution of individualOD405 values obtained by AgA-ELISA(IgG antibodies against T. cruzi pro-teins. Dotted line denotes cut-off.

ratio of positive test 33.69 (95% CI:20.10-56.47) and Likelihood ratio of nega-tive test 0.01 (95% CI: 0.00- 0.08). Therewere 1.12% (1/89) false negative and 2.94%(14/477) false positive results.

DISCUSSION

The control strategy for the elimina-tion of Chagas disease over the 1996-2010period is based on interruption of vectortransmission and the systematic screeningof blood donors (3). The interruption ofvectorial transmission has been obtainedwith great efficiency in vast areas of thesouth cone, but for financial reasons Cha-gas’ disease screening is not performed inall Latin-American blood banks, even by us-ing only one test (24); so we are far fromthe goal of a more accurate diagnose usingtwo tests. In consequence, there is an ur-gent need for an inexpensive test that canbe used in countries with low budgets andendemic for this disease.

To minimize false Chagas disease diag-noses, purified, recombinant or syntheticantigens have been evaluated. Most of themhave been found to be very specific buttheir sensitivity has not been good enoughfor their use in routine Chagas disease labo-ratory diagnosis. To improve this aspect,different combinations of them has beenevaluated (10, 15, 25, 26). This approachhas improved the sensitivity of the assaysbut has complicated their production andincreased their cost. The cost of productionof the ELISA test kits depends on numerousfactors like antigen preparation, conjugatesynthesis and purification, validation of re-agents, assay optimization and validation.We performed a preliminary assessmentstaking into account our individual costsand compared them with the most econom-ical Kit assay distributed in Venezuela. Inthis assessment our formulation is at leastten times less expensive.

In this study we present experimentalevidence that allows us to suggest that we

Vol. 51(1): 101 - 113, 2010

An inexpensive antigen for serodiagnosis of Chagas’ disease 109

TABLE II

DATA ON ASSESSMENT OF AGA-ELISA (n=566)

Epidemiology division, MSDA test

Positive Negative Total

AgA-ELISA

Positive 88 14 102

Negative 1 463 462

Total MSDS 89 477 566

TABLE I

CROSS REACTIVITY OBSERVED WITH AGA-ELISA AND THREE DIFFERENT COMMERCIAL KITS

CL/VL With other diseases (*)

AgA-ELISA 2.40/7.20 1.12

Bioschile 8.43/10.81 2.25

Pharmatest 22.89/21.62 12.36

Biokit 9.63/10.81 3.37

CL/VL: % of cross reactivity with Cutaneous Leishmaniasis/%of cross reactivity with Visceral Leishmaniasis.(*)Other diseases: Toxoplasmosis, Schistosomiasis, Giardiasis and Malaria.Serums positive for all 4 tests were considered mixed infected with T. cruzi, and did not were included in this Ta-ble.

can offer a very sensitive and specific test,that fulfills the financial requirements to beincluded in blood banks screening proce-dures as second diagnostic test for Chagas’disease, with high sensitivity and specificity.

Our antigen formulation (AgA), re-sulted from a very long experience of auto-claving epimastigote forms in stationaryphase of T. cruzi. The nutrient stress (16)increases the density of antigens recog-nized by the antiserum and/or induces theexpression of different antigenic moleculesof the parasite. This is particularly impor-tant because available kits are very effectiveat detecting blood donors presenting withhigh anti-T. cruzi antibody titers, but theresults are often questionable when the kitsare used for donors with low titers (27-29).Another factor that needs to be taken intoconsideration when one is using serologicaltests for Chagas’ disease is the cross-reactivity when Leihmaniasis cases are in-cluded, particularly visceral leishmaniasis(29). We present evidence that autoclavingpreferentially eliminated cross-reactive an-tigens. Consistent with observations in Fig.1, our formulation obtained the lowestcross reactivity results. It is remarkablethat with the all tests evaluated the higherpercentage of cross reactivity was observedin VL patients.

As an added benefit, the high tempera-ture and pressure reached with the auto-clave procedure eliminate, either by inacti-vation or by degradation, most of theproteolitic activity present in extracts ofthe parasite. It is important to highlightthat the effect that we obtained with T.

cruzi, cannot be obtained by the sametreatment with subcellular fractions fromthe parasite, and that this effect could bespecies specific, given that the same treat-ment does not work with Leishmania spp.(16).

A possible explanation to the effect in-duced by the combined treatment could be

that we obtain a preparation with high lev-els of specific antigens (suggesting thermo-stability) and low proportion of non-specificcross-reacting antigens (thermolabile). It ispossible that the lability of these non-spe-cific antigens is due to the fact that manyof them are conformational antigens, whilethe specific ones are stable or become moreexposed during the autoclaving process.Similar experiences have been observed inimmunocytochemestry (30, 31) where par-affin sections (5 mm) of the human intes-tines and ovaries were cut onto silane-coated glass slides, dewaxed with toluene,and dehydrated with serial ethanol solu-tions. The sections were autoclaved at120°C for 15 min in citrate buffer. The re-sults showed that autoclaving is necessaryfor antigen retrieval (30), or for localiza-tion of alpha estrogen receptor in mouse in-testine (31).

Regarding the 14 samples from en-demic areas tested with the AgA-ELISA thatdid not match the results from the epidemi-ological division of MSDS, there are twopossibilities: those samples are false posi-tive with our test, or they are true positivesthat the MSDS test did not see. At this mo-ment, in four patients from this group wewere able to perform a polymerase chain re-action and the results demonstrated thatthey were true positives and infected withT. cruzi, and not detected by the MSDStest.

We were able to demonstrate thatAgA-ELISA exhibits high sensitivity andspecificity levels, and low cross reactivity,comparable or superior to those shown bycommercial kits used in Venezuela. In addi-tion, serological diagnosis with AgA-ELISAhad almost a perfect agreement with theserological diagnosis of the epidemiologicaldivision of MSDS, suggesting that it couldbe used to perform blood bank diagnosis ofT. cruzi infection and to perform epidemio-logical studies with large numbers of per-

Investigación Clínica 51(1): 2010

110 Briceño y col.

sons, with good accuracy and low costs. Ad-ditional validations of this antigenic formu-lation in other countries and reassessmentsof commercial kits that are improving theirantigens as Pharmatest (Pharmatest Labo-ratories) are necessary.

ACKNOWLEDGEMENT

We want to thank the following re-searchers for the generous donation of hu-man sera: Dr. Marian Ulrich, Dr. MariaDiprisco, Dr. Belkisyole Noya and Dr. Oscar.Noya.

This work was partially supported byFONACIT G2000001530, LANPIP 200001639,LOCTI P935015 to and CDCH PI 09-35-5286-2007.

REFERENCES

1. Reporte del grupo de trabajo científico

sobre la enfermedad de Chagas, 2005.

TDR/GTC/06 http://www.who.int/tdr/pu-blications/publications/pdf/swg_chagas.pdf (open july 15th, 2008).

2. XVa Reunión de la Comisión Interguber-

namental del Cono Sur para la Elimina-

ción de Triatoma infestans y la Interrup-

ción de la Transmisión de Tripanosomia-

sis Transfusional (INCOSUR-Chagas).

2006. http://www.paho.org/Spanish/AD/DPC/CD/dch-incosur-xv.htm

3. Moncayo A, Ortiz Yanine MI. An updateon Chagas disease (human Americantrypanosomiasis). Ann Trop Med Parasitol.2006; 100: 663-677.

4. Brofen E, de Assis-Rocha FS, Machado

EB, Perillo MM, Romanha AN, Chiari E.

Isolation of Trypanosoma cruzi samples byxenodiagnosis and hemoculture from pa-tients with chronic Chagas’ disease. 1989.Mem Inst Oswaldo Cruz 84:237-240.

5. Potela-Lindoso A, Shikanai-Yasuda M.

Chronic Chagas disease: from xeno-diagnosis and hemoculture to polymerasechain reaction. Rev Saude Pub 2003; 37:107-115.

6. World Health Organization: Control of

Chagas disease. Who technical report Se-ries 1991; 811:38-47.

7. Ministerio de Saúde (Brasil): Normas téc-nicas para coleta, processamento e trans-fusão de sangue, componentes e derivados.Diário oficial da união, 11 August 1989;portaria 721.

8. Sáez-Alquézar A, Murta M, Pereira

Marques W, Rodrigues da Silva G. The re-sults of an external quality control pro-gram for serological screening for antibod-ies against Trypanosoma cruzi in blood do-nors in Brazil. Rev Panam Salud Publica2003; 13(2/3):129-137.

9. Gadelha AA, Verçosa AF, Lorena VM,

Nakazawa M, Carvalho AB, Souza WV,

Ferreira AG, Silva ED, Krieger MA,

Goldenberg S, Gomes YM. Chagas’ dis-ease diagnosis: comparative analysis of re-combinant ELISA with conventional ELISAand the haemagglutination test. Vox Sang2003; 85:165-170.

10. Silveira F, Umezawa E, Luquetti A. Cha-gas’ disease: recombinant Trypanosomacruzi antigens for serological diagnosis.Trends Parasitolol 2001; 17:286-291.

11. Roddy P, Goiri J, Flevaud L, Palma PP,

Morote S, Lima N, Villa L, Torrico F,

Albajar-Viñas P. Field evaluation of a rapidimmunochromatographic assay forTrypanosoma cruzi infection using wholeblood in Sucre, Bolivia. J Clin Microbiol2008; 46:2022-2027.

12. Meira WS, Galvão LM, Gontijo ED,

Machado-Coelho GL, Norris KA, Chiari E.

Trypanosoma cruzi recombinant comple-ment regulatory protein: a novel antigenfor use in an enzyme-linked immuno-sorbent assay for diagnosis of Chagas’ dis-ease. J Clin Microbiol 2002; 40:3735-3740.

13. Umezawa ES, Luquetti AO, Levitus G,

Ponce C, Ponce E, Henriquez D, Revollo

S, Espinoza B, Sousa O, Khan B, da

Silveira JF. Serodiagnosis of chronic andacute Chagas’ disease with Trypanosoma

cruzi recombinant proteins: results of acollaborative study in six Latin Americancountries. J Clin Microbiol 2004; 42:449-452.

Vol. 51(1): 101 - 113, 2010

An inexpensive antigen for serodiagnosis of Chagas’ disease 111

14. Silveira-Lacerda EP, Silva AG, Junior SF,

Souza MA, Kesper N, Botelho-Filho A,

Umezawa ES. Chagas’ disease: applicationof TESA-blot in inconclusive sera from aBrazilian blood bank. Vox Sang 2004; 87:204-207.

15. Luquetti AO, Ponce C, Ponce E,

Esfandiari J, Schijman A, Revollo S, Añez

N, Zingales B, Ramgel-Aldao R, Gonzalez

A, Levin MJ, Umezawa ES, Franco da

Silveira J. Chagas’ disease diagnosis: amulticentric evaluation of Chagas Stat-Pak, a rapid immunochromatographic as-say with recombinant proteins ofTrypanosoma cruzi. Diagn Microbiol InfectDis 2003; 46:265-271.

16. Campos Y, Briceño L, Reina K, Figarella

K, Pérez JL, Mosca W. Serological diagno-sis of Chagas disease: evaluation and char-acterisation of a low cost antigen withhigh sensitivity and specificity. Mem InstOswaldo Cruz 2009; 104:914-917.

17. Mosca W, Briceno L, Hernandez MI. Cellmediated immunity in Chagas’ disease.Trypanosoma cruzi antigens induce sup-pression of the in vitro proliferative re-sponse of mononuclear cells. Mem InstOswaldo Cruz. 1991; 86:147-152.

18. Peralta Soler A, Gilliard W, Xiong G,

Knudsen K, Martin-Suarez J, Mota- Gam-

boa J, Mosca W, Briceño-Zoppi L. Over-expression of neural cell adhesion mole-cule (ncam) in Chagas’ myocarditis. Hu-man Pathology 2001; 32:149-155.

19. Briceño L, Rodríguez EM, Medina M, Pa-

redes A, León G, Hung M, Walter M.

Enfermedad de Chagas: evaluación de es-tuches comerciales para el diagnóstico se-rológico. Arch Hosp Vargas 2003; 45:35-42.

20. Bangdiwala SI, Muñoz SR. Measurementof reliability and validity of clinical instru-ments. Rev Med Chil 1997; 125:466-473.

21. Sim J, Wright CC. The kappa statistic inreliability studies: use, interpretation andsample size requirements. Phys Ther 2005;85:257-268.

22. Landis JR, Koch GC. The measurement ofobserver agreement for categorical data.Biometrics 1977; 33:159-174.

23. Mackinnon A. A spreadsheet for the calcu-lation of comprehensive statistics for theassessment of diagnostic tests and inter-rater agreement. Computers in Biologyand Medicine 2000; 30:127-134.

24. Pinto JC. Southern Cone Initiative for theelimination of domestic populations of Tri-atoma infestans and the interruption oftransfusional Chagas disease.Historical as-pects, present situation, and perspectives.Mem Inst Oswaldo Cruz 2007; 102:11-88.

25. Ferreira AW, Belem ZR, Lemos EA, Reed

SG, Campos-Neto A. Enzyme-linkedimmunosorbent assay for serological diag-nosis of Chagas’ disease employing aTrypanosoma cruzi recombinant antigenthat consists of four different peptides. JClin Microbiol 2001; 39(12):4390-4395.

26. Chang CD, Cheng KY, Jiang LX, Salbilla

VA, Haller AS, Yem AW, Bryant JD,

Kirchhoff LV, Leiby DA, Schochetman G,

Shah DO. Evaluation of a prototypeTrypanosoma cruzi antibody assay with re-combinant antigens on a fully automatedchemiluminescence analyzer for blood do-nor screening. Transfusion 2006; 46:1852-1859.

27. Marcon GE, Andrade PD, de Albuquerque

DM, Wanderley Jda S, de Almeida EA,

Guariento ME, Costa SC. Use of a nestedpolymerase chain reaction (N-PCR) to de-tect Trypanosoma cruzi in blood samplesfrom chronic chagasic patients and pa-tients with doubtful serologies. DiagnMicrobiol Infect Dis 2002. 43:39-43.

28. Silveira-Lacerda EP, Silva AG, Favoreto

S, Souza MA Jr. Kesper N, Botelho-Filho

N, Umezawa ES. Chagas disease: applica-tion of TESA-blot in inconclusive sera froma Brazilian blood bank. Vox Sang 2004;87:204-207.

29. Caballero ZC, Sousa OE, Marques WP,

Saez-Alquezar A, Umezawa ES. Evaluationof serological tests to identify Trypano-

soma cruzi infection in humans and deter-mine cross-reactivity with Trypanosoma

rangeli and Leishmania spp. Clin VaccineImmunol 2007; 14:1045-1049.

30. Ehara H, Koji T, Deguchi T, Yoshii A,

Nakano M, Nakane PK, Kawada Y. Expres-

Investigación Clínica 51(1): 2010

112 Briceño y col.

sion of estrogen receptor in diseased hu-man prostate assessed by non-radioactivein situ hybridization and immunohisto-chemistry. Prostate 1995; 27:304-313.

31. Kawano N, Koji T, Hishikawa Y, Murase

K, Murata I, Kohno S. Identification and

localization of estrogen receptor alpha-and beta-positive cells in adult male andfemale mouse intestine at various estrogenlevels. Histochem Cell Biol. 2004; 121:399-405.

Vol. 51(1): 101 - 113, 2010

An inexpensive antigen for serodiagnosis of Chagas’ disease 113